Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

A high performance Open API platform for disaster management, integrating UAVs, mobile and IOT devices

Kyritsis Sarantis

Full record

---

Summary

Search and rescue (SAR) constitutes a crucial, recurrent and integral challenge for civil protection entities. With the use of current technology, multiple parts of a SAR mission can be inter-connected in real-time, by using multiple aspects from the world of GIS systems, as well as leveraging the capabilities of embedded systems for on the fly application deployment.Multiple approaches have been undertaken so far towards bridging these fields, but the proposed system is utilizing the advantageous elements of primary memory databases, as well as the performance of statically typed compiled languages, such as Golang.The goal of this thesis is the creation of an Open API, real time and highly performant disaster management system. The key elements of this system are that it is easily deployable, easily configurable and implements a variety of different communication protocols. The drive for this implementation is that the resulting system can be deployed on any Linux-based embedded system, works with a variety of messaging protocols and is completely agnostic of the underlying network infrastructure. For the creation of this platform, a variety of technologies have been used. The deployment system is based on docker containerization, the programming language is Golang, while the main database used is Redis.The resulting platform has been thoroughly stress tested in handling loads of transactions, simulating high load real world scenarios. Through this platform, multiple field agents (i.e. drones, rescuers, volunteers, teams of people) can coordinate their actions by utilizing whatever available network infrastructure is still operating, or can be easily deployed on the spot. Moreover, people that are facing problems (i.e. cannot move, are injured, or are trapped) can directly send their position, as well as their status through this platform, so that the responding rescuers can approach them efficiently. Performance tests indicate that the system can handle 33269 concurrent requests per second regarding creations and updates of data in the system and 57543 concurrent requests per second fetching data from the platform. To put things into perspective an ArcGIS Server REST 2D Vector mapping services supports up to 79,710 requests per hour, and a heavily optimized PostGIS installation can reach up to 7432 requests per second.The final system offers a real-time disaster management platform, heavily optimized for fast messaging performance between the users of the platform, and mainly oriented towards team coordination in SAR operations.

Available Files

Services